Design, construction and test of a surgical keratometer

346 ARQ. B R A S . OFTAL . 6 1 ( 3 ) , J U N HOI J 998 SUMMARY . . ÇJbjec,t{ye: The preserit study aims to draw, build andt� 'and l�w (;Qst surgiC�l i<�t.âtometer. Mêt����:·Th.é.pd�c���i! described by Plácido da Costa w�� � co.�struct the prototype. A circular ring of light was . ·bHi.� attacited.. l �9 a surgical nticroscope. The zoom system présent in titi! t microsc0pe'provided � bigger o r snúíller image o f the riôg to th observerllts size wôuld th.éô be conípared to a circular reticl�IôÇâ irisideth� êicuh1r lens ofthe níicroscope�N umeric values �ere�ç��� . by a gr�duàted scale. The tlattest and steepest meridians . were �l� located. . · �··


INTRODUCTION
Curiosity about the function and fo rm ofthe comea dates fr om the year 150 BC. Galen, at that time, already depicted the eye with a curved form of its external layer endowing it with optical properties 1 • However, refraction laws were only detailed and applied to vision by Kepler, in 1619, ten years after Christopher Scheiner had measured for the first time the curvature of the cornea when comparing reflections o f a window, on a sphere o f known size, with those produced by the comea 2• In 1 854, Hermann von Helmholtz applied part o f Scheiner' s principies to the first device destined to measure the radius o f the curvature o f the cornea, o r keratometer 3 .
In the XIX century, measurements of the cornea were used in the attempt to determine the dioptric power of the eye and Alvar Gulstrand presented his schematic eye. With the universalization ofthe use of contact lenses, keratometry gained another impulse and the global form of the cornea started to be restudied.
As reported by Bicas 4, until 1941 there were only two general articles published, by Gama 5 and by Prado 6 , on this matter in the Brazilian literature. However, in 1967, a devi c e able to measure the curvature o f the comea, the depth ofthe anterior chamber and the visual field was presented.
The device, designed and constructed by Bicas represented a landmark in the history o f biomedical engineering applied to ophthalmoloqy in Brazil 4.
After 1970 refractive comeal surgery was introduced in the US and showed to be a decisive advance in the search for the form of the comea in view of its possible surgical modification.
The determination of the comeal curvature during ophthal mologic procedures was introduced in 1962 in Colombia by lgnacio Barraquer 7 and fr om then on severa! instruments for this purpose proliferated. Surgi cal keratometers have wrongly been divided into quantitative and qualitative. Wrongly, because the word kerato (from the Greek comea), meter (from the Greek measure) implies quantitative measurement and therefore, "qualitative keratometers" should be called keratoscopes.
While precision obtained with a keratoscope regarding evaluation of the comeal curvature at its different meridians (astigmatism) does not exceed 2.0 D 8, keratometers present a precision up to O .1 O D 9. The simp I e use and low cost lead to proliferation o f surgi cal keratoscopes but their limited precision opens the way fo r a simple and low cost surgical keratometer.
The purpose o f this study is to design, construct and test a simp le low cost keratometer.

METHODS
According to the principie described by Plácido 9 , a circu lar illuminating system was idealized and attached to the obj ective lens of a surgical microscope presenting a fo cal distance of 175 mm. Fluorescent light was used for the illumination (25 W daylight, Osram, USA), coupled to a metal support. The lower plate of the support was hollowed and covered with a transparent mask where a circular ring was printed. The illuminated ring had a 133 mm diameter and an approximately 1 mm width.   The purpose of such an illuminating system was to present a circular image to a polished convex surface, thus producing a virtual, direct and smaller image, located at approximately 2 mm fr om its posterior surface. The image would be observed and compared with a reticle located at the ocular lens of the microscope. Such reticle presented a double circular line (360 degrees) circled by a semi-compass subdivided into 5 are degrees.
With the varifocal magnifying (zoom) system present in the microscope, the observer could increase or decrease the observed field and, consequently, the image produced by the surface. The purpose o f magnifying, o r decreasing, the image was to make the image ofthe surface coincide with the circular reticle ofthe microscope. Ifthere is no astigmatism, the image produced would be circular and the superposition complete. lf there is astigmatism two points of coincidence would that required in order to define the most steep and flat meridians o f the surface (Fig. 1).
A scale was coupled to the manual (zoom) magnifying system for the determination of the numerical values. This scale was empirically constructed measuring the curvature radii o f steel spheres o f known size.
The device was tested by means of evaluation by three observers of the astigmatism present in eight toric plastic lenses (with astigmatism at the anterior face) oriented in three different positions (90, 180 and 45 degrees). The lenses presented respectively O. 74, 0.98, 1.22, 1.60, 1.96, 2. 76, 3.86 and 5.64 D mean astigmatism as measured by the comeal videokeratograph (EyeSys, TX, USA).
In order to determine precision of the astigmatism meridian, five toric lenses (with 0.74, 1.22, 1.60, 1.96 and 3.86 D mean astigmatism) were inclined in preestablished positions and measured by three observers (Fig. 2). Orienta-

RESULTS
The device was idealized, assembled and empirically calibrated (Fig. 3).
Regarding the quantification of astigmatism, there was a statistically significant reproducibility (p < 0.005) in the three e.

DISCUSSION
Siroplicity o f construction, using sirople optical principies, and roechanisms already present in the roicroscope are reflected in cost and use o f the prototype. The cost remained well below that o f the conventional keratometers and its use was accepted by the three observers. In spite o f its simplicity, and adjustments required for precision ofthe printed scale, the results shown are encouraging.
Surgical keratoscopes which do not present an externa! reference are precise up to a maximum of 2.5 D 1 1 • We could speculate about the fa ct of using the Gestalt theory which considers that there is a tendency of the huroan being to turn irregular fo rros into regular fo rms, thus "nullifying" slightly elliptic fo rros, which would correspond to low astigroatisros 1 2 . On coroparing an iroage with an externa! reference, a higher precision ofthe astigmatism value is attained, however without determining the absolute value of the steepest and flattest  Graph 2 • Angular erro r present on measuring lenses with increasing astigmatism. Astigmatisms higher than 1.60 showed standard deviations lower than 6.3 degrees.
meridians. Commercial keratometers cots between US$ 14,000 and 29,000 and present a precision ranging from 0.10 to 0.25 D 9, but more specific data on their performance are scarce, there being a literature regarding only clinicai results obtained by their use. The tested model has a precision ranging from 0.29 to 0.62 D and an approximate cost of US$ 500,00. Clinicai tests with anesthetized eyes are required fo r clinicai comparison.
The lack o f statistical significance regarding reproducibility of astigmatism reading in 1.60 and 1.96 D range in the three positions by the three observers is not easily explained. Technical limitations in the printing ofthe reading scale which would not reflect a correct curvature value, or imperfections of these lenses, could help but are not the last word regarding this fa ct. These fa ctors should be accounted for when constructing a prototype for clinicai use.
Concerning orientation o f the meridians, comparison with commercial devices is again difficult, since values with a precision o f 1 O degrees, described by Frantz et al. 13, in regard to the keratometer produced by the Nidek company (Paio Alto, CA, USA), render implicit to the reader a constant precision in ali astigmatism values. The keratometer presented in this study has a precision of approximately ± 10 degrees ofinclination in astigmatisms higher than 1.60 D, and of approximately 6 degrees in astigmatisms higher than 1.96 D. Such a precision does not characterise the de vice as being useful, for example, in present-day photo-refractive surgeries, but in surgeries related to higher astigmatisms, such as extracapsular cataract extrac tions, corneal resections and come a keratoplasty. In these cases there is detection and measurement of astigmatism with characterization o f the steepest meridian.